Amplification of neurotoxic HTTex1 assemblies in human neurons

Abstract

Huntington’s disease (HD) is a genetically inherited neurodegenerative disorder caused by expansion of a polyglutamine (polyQ) repeat in the exon-1 of huntingtin protein (HTT). The expanded polyQ enhances the amyloidogenic propensity of HTT exon 1 (HTTex1), which forms a heterogeneous mixture of assemblies with a broad neurotoxicity spectrum. While predominantly intracellular, monomeric and aggregated mutant HTT species are also present in the cerebrospinal fluids of HD patients, however, their biological properties are not well understood. To explore the role of extracellular mutant HTT in aggregation and toxicity, we investigated the uptake and amplification of recombinant HTTex1 assemblies in cell culture models. We find that small HTTex1 fibrils preferentially enter human neurons and trigger the amplification of neurotoxic assemblies; astrocytes or epithelial cells are not permissive. The amplification of HTTex1 in neurons depletes endogenous HTT protein with non-pathogenic polyQ repeat, activates apoptotic caspase-3 pathway and induces nuclear fragmentation. Using a panel of novel monoclonal antibodies and genetic mutation, we identified epitopes within the N-terminal 17 amino acids and proline-rich domain of HTTex1 to be critical in neural uptake and amplification. Synaptosome preparations from the brain homogenates of HD mice also contain mutant HTT species, which enter neurons and behave similar to small recombinant HTTex1 fibrils. These studies suggest that amyloidogenic extracellular mutant HTTex1 assemblies may preferentially enter neurons, propagate and promote neurodegeneration.

1. Introduction

Huntington’s disease (HD) is an autosomal genetically inherited neurodegenerative disorder characterized by debilitating motor, psychiatric, and cognitive symptoms (Bates et al., 2015, Gosh and Tabrizi, 2018). Expansion of a CAG repeat (>35) in the exon 1 of huntingtin HTT gene, which translates into an abnormal polyglutamine (polyQ) tract, is the underlying cause of Huntington’s Disease (HD). (The Huntington’s Disease Collaborative Research Group, 1993). Mutant HTT exon-1 (HTTex1) released by the enzymatic cleavage of full-length protein and/or by aberrant splicing of mutant HTT mRNA, is the most aggregation-prone species, accumulates in the brains of HD patients and is sufficient to induce severe HD-like pathology in animal models (DiFiglia et al., 1997, Davies et al., 1997, Lunkes et al., 2002, Sathasivam, et al., 2010, Bates et al., 2015, Neueder et al., 2018, Yang et al., 2020). The aggregation of mutant HTTex1 produces a heterogeneous mixture of assemblies including oligomers, fibrils and inclusion bodies, which may have different biological properties (Arrasate et al., 2004, Kim et al., 2016, Sahoo et al., 2016, Bäuerlein et al., 2017). For example, transient expression of mutant HTTex1 in culture models revealed that the accumulation of soluble oligomeric species coincides with neurotoxicity (Arrasate et al., 2004, Nucifora et al., 2012). On the other hand, fibrils of HTTex1 may interact with cellular membranes and disrupt their architectures and vital functions (Bäuerlein et al., 2017). The promiscuous interaction of misfolded mutant HTTex1 with cellular proteins may also contribute to the heterogeneity of assemblies and toxicity spectrum (Kim et al., 2016, Wanker et al., 2019).

The oligomerization of mutant HTTex1 in vitro occurs by a stepwise mechanism and is influenced by several factors such as interaction with biological membranes and the formation of an oligomeric seed structure, which acts as a nucleating agent and accelerates aggregation (Pandey et al., 2017, Tao et al., 2019). Seed structures may also assemble in vivo, however, little is known about their structures and roles in aggregation and toxicity. Seeding-competent mutant HTT assemblies capable of promoting the aggregation of monomeric HTTex1 in vitro and in cell models have been isolated from the brains of HD animal models and postmortem brain homogenates and cerebrospinal fluids (CSF) of HD patients. In these studies, the levels of mutant HTT seeds positively correlated with disease severity in the HD patients (Morozova et al., 2015, Ast et al., 2018, Lee et al., 2020). Interestingly, similar to other amyloidogenic proteins such as α-synuclein and Tau, mutant HTTex1 may have the propensity to propagate by a prion-like mechanism (Jucker and Walker, 2018, Pearce and Kopito, 2018, Masnata et al., 2019). For example, mutant HTTex1 assemblies spread between neurons in the brains of Drosophila models of HD and form aggregates by recruiting monomeric HTTex1 (Pearce et al, 2018). Moreover, cerebrospinal fluids (CSF) of HD patients seed mammalian cells engineered to express mutant HTTex1-EGFP and induce protein aggregation (Tan et al., 2015, Lee et al., 2020). These encouraging findings provide a new direction in HD research and may have implications for the spread and propagation of neurotoxic mutant HTT assemblies in HD.

The neurodegenerative aspect of HD may release various mutant HTT assemblies in the circulation and CSF. Moreover, HTT is actively and passively secreted from cultured cells and neurons in the HD animal models (Trajkovic, et al., 2017, Caron et al., 2021). Indeed, the levels of mutant HTT in plasma and CSF of HD patient are being used as biomarkers evaluating the efficacy of therapeutics in HD patients (Tabrizi et al., 2019). While few recent studies demonstrate that extracellular amyloidogenic HTTex1 fibrils have the propensity to enter cell lines and neurons in mouse models (Masnata et al., 2019, Lee et al., 2020), the biological properties of extracellular mutant HTT seeds including the mechanism for cell entry, amplification, and neurotoxicity remain to be investigated. Here, we performed a structural and biological characterization of small HTTex1 fibrils. We report that small HTTex1 fibrils preferentially enter human neurons, accumulate in the nucleus, amplify and produce neurotoxic assemblies. We further demonstrate that conformations within the N-terminal 17 amino acids and proline-rich domains (PRD) of HTT are critical for the neural entry and amplification. We propose that extracellular assemblies of HTTex1 may contribute to amplification of neurotoxic assemblies.

2. Materials and Methods

2. 1. Antibodies

PHP1-PHP3 mouse monoclonal antibodies (mAbs) were reported previously (Ko et al., 2018). The new PHP5 and PHP6 mAbs were isolated from a mAb library made to the N-17 peptide with 7 glutamines at the C-terminus, and PHP7-PHP10 were produced to sonicated mutant HTTex1 fibrils (Khoshnan et al., in preparation). Clones were selected from each hybridoma library for binding to HTT species by ELISA, Western blots and dot blots as previously described ((Ko et al., 2018), (Supplementary fig. 2A and B). Antibody to beta III Tubulin was from Abcam (Cat# ab18207, 1:1000). The secondary antibodies were goat anti-rabbit Alexa Fluor 488 Cat# A32731, goat anti-mouse Alexa Fluor 488, Cat# A28175, goat anti-mouse Alexa Fluor 594 Cat# A32742, Life Technologies, 1:1000 and goat anti-mouse, Horseradish Peroxidase (HRP) Cat# A16072, Invitrogen, 1:10000.

2.2. Dot blot assay

A strip of PVDF membrane was pre-wet in 100% methanol for 15 s, soaked in water for 2 min and equilibrated for 5 min in TBS-T (0.05 % Tween, pH 7.4). A sheet of Whatman filter paper was then soaked in TBS-T and placed on a dry sheet of Whatman filter paper on top of some dry paper towels. The PVDF membrane was placed on top of filter stack and 2 μL of each protein was spotted on a pre-marked grid. The membrane was dried to fix proteins to it for 1.5 h at room temperature. The membrane was then blocked, probed with each indicated primary antibody followed by treatment with goat-anti-mouse HRP secondary antibody, and detected by the addition of chemiluminescent agent Clarity™ Western ECL Substrate (Cat# 1705060, Bio-rad).

2. 3. Engineering of recombinant PHP2 antibody and lentivirus production

The DNA fragments encoding the antigen binding domain of VH and VL of PHP2 were amplified from a cDNA library made to mRNA of parental hybridoma by standard PCR methods using mixture of commercial primers and sequenced (Khoshnan et al., 2002). The amplified cDNAs were assembled into an IgG2 backbone (provided by Dr. Alejandro Balazs at the Ragon Institute of MGH, MIT and Harvard) by Gibson assembly (New England biolabs) and subsequently cloned into a lentiviral vector (FUGW). Control and PHP2 recombinant viral particles were produced in HEK-293 cells as described previously (Khoshnan and Patterson, 2012). Viral titers were determined using a GFP lentivirus as a reference. Subsequently, MESC2.10 neural progenitor cells (NPCs) were transduced at multiplicity of (2:1). Supernatant of neurons form the transduced NPCs were tested for antibody secretion and PHP2 antibody binding to HTT fibrils was confirmed by Western blots.

2.4. Purification of thioredoxin (TRX)-HTTex1 fusion protein

Expression and purification of wild-type (Q25) and mutant (Q46) HTTex1 fusion proteins has been described previously (Bugg et al., 2012, Isas et al., 2015, Isas et al., 2017). Briefly, bacterial cell pellet containing the expressed fusion protein (6XHis-TRXA-HTTex1/or HTTex1–111C variant for labeling) were lysed and cell debris were removed by centrifugation. The clarified lysates containing the fusion protein were loaded onto Ni-column (NiHis60 super flow resin) and washed with a low concentration of imidazole saline buffer. The fusion protein was eluted with high concentration of imidazole saline buffer. The HTTex1–111C variant was first labeled with AlexaFlour 555 (1 to 3 molar ratios for 3hr), diluted 1 to 10 with low salt buffer, loaded on to anionic exchange resin (Mono Q), and further fractionated with a linear gradient (25mM to 1M salt) to remove free label and elute the labelled Alexa sample.

2.5. Preparation and purification of HTTex1 fibrils

Monomer HTTex1 (Q46) ((HTTex1) (Q46_111Alexa555)) was produced, by the removal of the TRX fusion tag enzymatically (EKMax), followed by separation on reverse phase column (C4) with an acetonitrile gradient as previously described (Pandey et al., 2018). HTTex1 fibrils or (Alexa555 labelled fibrils) were made by first solubilizing lyophilized protein powder from previous step in 0.5% TFA (v/v) in methanol, determined the concentration, and removed the organic solvent by gentle N₂ gas stream, resulting in a thin clear protein film. The protein film was resuspended in ice cold buffer (20 mM Tris pH7.4, 150 mM NaCl) and adjusted to between 20 to 25 μM concentration. Fibrils reaction started by adding 1% (molar ratio) of sonicated HTTex1 seeds as previously described (Isas et al., 2017) and incubated at 4°C overnight. In a separate reaction, 20% Alexa labelled fibrils were made by adding 4 to 5 μM of HTTex1_111alexa555 solubilized monomer to the Httex1 unlabeled sample. After overnight incubation fibrils were harvested by ultra-centrifugation, and the resulting translucent pellets were resuspended in TFA:H20 (1:4000) at a concentration of 1mg/ml and fragmented using sonication. Mutant HTTex1 in which each residue in PRD domain was replaced with a Pro residue (HTTex1mL17) was generated for probing the binding site of PHP1 and PHP2 as recently reported (Ko et al., 2018). HTTex1mL17 Fibrils were made as described for the HTTex1 in the section above.

2. 6. Treatment of cells with HTTex1 seeds

Neurons were derived from embryonic human MESC2.10, IPSC-derived neuronal progenitor cells (NPCs) (provided by Dr. Alysson Muotri, UCSD) according to standard protocols (Khoshnan et al., 2012). Astrocytes were derived from IPSCs by culturing progenitors in DMEM/F12 supplemented with N2 growth factor and 0.5% FBS (Cat# ES-009-B, Millipore) for 14–20 days. Caco-2 and Neuro2A were obtained from ATCC and cultured according to instructions provided. Each line was treated with 1.25 nM, 5 nM or 10 nM of sonicated HTTx1 fibrils for the indicated time points in the figure legends and processed for immunocytochemistry and confocal microscopy as described previously (Khoshnan, et al., 2012). For antibody inhibition assay, 10 nM of sonicated fibrils were pre-incubated with 1 μg of each antibody clone indicated in the figures for 2 hr at RT and subsequently added to growth medium and processed as above.

2.7. In vitro seeding assay with HTT species

Sonicated HTTx1 seeds (10 ng) were incubated with recombinant WT HTTex1 monomers or 100 μg of total protein from human neural lysates for 4 hr at 25°C with continuous agitation. For examining the seeding capability of amplified products, 5 % (v/v) of seeded neuronal lysate was incubated with 100 μg of fresh lysate for 4 hr at 25°C. Semi-denaturing detergent agarose gel electrophoresis (SDD-AGE) was performed to examine the amplified products and the corresponding seeds as described previously (Halfman and Lindquist, 2008) with some modifications. Briefly, seeded lysates were loaded in a 1.5% agarose gel in 1X Tris-acetate-EDTA (TAE) (Tris base, glacial acetic acid and EDTA) containing a final concentration of 0.1% SDS. Using 1X Tris-buffered saline (TBS), downward capillary action was used to transfer protein to immune-blot PVDF membrane (Merck cat# IPVH00010). Membranes were blocked with 5% milk in wash buffer (0.05% Tween in PBS) and incubated with the indicated primary anti-HTT antibodies overnight at 4°C. HRP-conjugated goat anti-mouse secondary antibody (1:10000) in blocking buffer was then applied for 2 h and developed with ECL substrate and X-ray film.

2. 8. Western blot analysis of seeded neurons

MESC2.10 derived neurons grown in 10 cm culture plates were treated with 10 nM of sonicated HTTex1 fibrils for 24 hr. The seeded neurons were harvested and lysed in RIPA buffer (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% NP-40, 0.1% SDS and protease inhibitor). 50 μg of neuronal lysate and corresponding mHTTex1 seed were boiled for 5 min with sample loading buffer and loaded into a 4–20 % polyacrylamide gel (SDS-PAGE, Criterion Bio-Rad) to detect monomeric endogenous HTT or by semi-denaturing detergent agarose (1.5%) gel electrophoresis (SDD-AGE) to probe for the seeded aggregated products. Different antibodies indicated in the figures and figure legends were used for detecting different species of HTT.

For proteinase K resistance assay, 0.1 μg of mHTTex1 seed or 50 μg of seeded neuronal lysate was incubated with varying doses of proteinase K (0 to 0.5 μg/mL, Qiagen) for 30 min at 25°C and heat inactivated at 75°C for 10 min. The digested products were analyzed by SDD-AGE and western blot. PHP2 was used as the detection antibody.

2.9. Immunodepletion of HTT in neural lysates

Immunodepletion of HTT was performed by incubating100 μg of neuronal lysates with 2 μg of each anti-HTTex1 or control antibodies overnight at 4°C with rotation. The antibody-HTT complexes were precipitated with protein G covalently conjugated to magnetic beads (Thermo Scientific, P188803). The HTT-depleted lysates were seeded with 10 ng of sonicated HTTex1 fibrils for 4 hr at 25°C with rotation. SDD-AGE was performed to detect the amplified products. The primary antibody used for western blot analysis was PHP1, which is reactive to mutant HTT aggregates in HD animal models (Ko et al., 2018).

2.10. Sucrose fractionation of HD brains

Synaptosome and ER/Golgi fractions were prepared from individual mouse forebrains according to a recently published method (Mastro et al., 2020). Briefly, forebrains of 9-month old ZQ175 HD mice were dissected from each animal, rinsed in Buffer A (0.32 M sucrose, 1 mM NaHCO₃, 1 mM MgCl₂, 0.5 mM CaCl₂, 0.1 mM phenylmethylsulphonyl chloride (PMSF, Sigma Millipore, St. Louis, MO). Each individual forebrain was homogenized in Buffer A (10% w/v, 4.5 ml for mice) with 12 up and down strokes of a Teflon/glass homogenizer at 900 rpm. Homogenates were subjected to centrifugation at 1400g for 10 min. The pellet was resuspended in Buffer A to 10% w/v (3.8 ml), homogenized (three strokes at 900 rpm) and subjected to centrifugation at 710 g for 10 min. The final resultant pellet (P1) was harvested as a crude fraction containing the nuclei. The two supernatants (S1) were combined and subjected to centrifugation at 13,800g for 10 min. The resulting pellet (P2) was resuspended in Buffer B (0.32 M sucrose, 1 mM NaHCO₃; 2 ml for mice), homogenized with 6 strokes at 900 rpm in a Teflon/glass homogenizer, and layered onto a discontinuous sucrose gradient (equal parts 0.32M, 0.85 M, 1.0 M, and 1.2 M sucrose in 1 mM NaH₂CO₃ buffer (10.5 ml). Gradients were subjected to centrifugation for 2 hours at 82,500g in a swinging bucket rotor. The bands between 0.32M and 0.85M sucrose (light membranes, Myelin), 0.85M and 1.0M sucrose (light membranes, Endoplasmic Reticulum, Golgi), 1.0M and 1.2M sucrose sections (Synaptosomes) were collected.

2.11. Assessment of seeding activity of protease-resistant brain fractions

Synaptosome and ER/Golgi fractions isolated from ZQ175 mouse brain by sucrose fractionation were dialyzed using Slide-A-Lyzer Dialysis Cassette (Thermo Fisher Scientific, 66203). The dialyzed mouse brain fractions isolated by sucrose fractionation were incubated with 0.01 μg/mL of Proteinase K at 37°C for 1 hr and heat inactivated at 75°C for 10 min. MESC2.10 neurons in 12-well plate were then incubated with 50 μg of each fraction at 37°C for 24 hr. This was followed by immunocytochemistry to examine for the presence of seeding-competent fibrils and assembly formation.

2.12. Immunocytochemistry and nuclear damage quantifications

Cells were fixed with 4% formaldehyde in PBS at room temperature for 30 min. After permeabilization and blocking (70% methanol in PBS, at least 1h at −20°C, 10% normal goat serum and 2% BSA in PBS, 30 min at room temperature); cells were incubated overnight with specific primary antibodies at 4°C. After washing in PBS, cells were incubated with appropriate AF 594 or 488 conjugated secondary antibodies in blocking solution for 2 hr at RT, washed and mounted in Vectashield with DAPI. At least 6 random pictures were captured using a Leica SP8 confocal microscope and analyzed using LAS X software. For each picture, the total number of intact cells as well as those with fragmented nuclei were counted. The percent of neurons with intracellular HTT aggregates (seeded neurons) and neurons with nuclear fragmentation was then calculated for each condition and plotted.

2.13. Caspase-3 Assays

Neurons were treated with 10 nM of sonicated HTTex1 seeds for 24 hr. PHP1 antibody (1:1000) was used to detect intracellular HTT fibrils and caspase-3 activity in untreated and treated neurons was monitored by immunocytochemistry using anti-caspase-3 antibody (Abcam, Cat# 32351, 1:1000). Images were captured using a Leica SP8 confocal microscope and analyzed by LAS X software. Neurons with aggregates and active caspase-3 were quantified as described in the section above.

2.14. Statistical analysis

For pairwise comparisons, Student’s t-test was used and analysis of variance (ANOVA), followed by the Tukey’s post hoc test was performed for multiple comparisons. P values <0.05 were considered as significant and for all results, values are presented as the mean and error bars indicate standard error of the mean (SEM). Statistical significance was assigned as ∗ = P < 0.05, ∗∗ = P < 0.01, ∗∗∗ = P < 0.001.

3. Results

3.1. HTTex1 assemblies preferentially enter human neurons and amplify

To explore the biological properties of recombinant HTTex1 (46Qs), we incubated pre-assembled fibrils (Isas et al., 2017) with human neurons and examined for uptake and amplification. Mutant HTTex1 fibrils bound to neural surfaces but lacked any detectable seeding activity. However, sonication of the fibrils generates smaller assemblies (HTTex1 seeds) (Supplementary fig. 1), which enter neurons derived from an embryonic neuronal progenitor cell line (MESC2.10) and produce assemblies detected by antibodies specific to HTTex1 fibrils. The amplification of HTTex1 in neurons is progressive affecting ~60% of the cells by 24 hr (Fig. 1B). To verify the amplification of HTTex1 seeds, we examined equivalent ratios of the seeds and lysates of seeded neurons by semi-denaturing detergent agarose gel electrophoresis (SDD-AGE) followed by Western Blotting (WBs) and find that the amplified HTTex1 seeds are a heterogeneous mixture of assemblies larger than ~ 250 kDa (Fig. 1C). Moreover, the assay is relatively sensitive since nanomolar concentrations of the HTTex1 seeds are sufficient for amplification (Supplementary fig. 2). Similar to recombinant HTTex1 fibrils, the amplified products in neuronal lysates are partially resistant to proteinase-K (PK) digestion further supporting the formation of complex aggregates (Fig. 1D and E, respectively).

Figure 1.

HTTex1 assemblies enter and amplify in neurons. (A) Human neurons derived from MESC2.10 embryonic stem cell line was treated with mutant HTTex1 fibrils or sonicated fibrils (seeds), incubated for the indicated time points and processed for ICC using anti-PRD PHP1 antibody (HTT). Insert in the bottom right panel is a magnified seeded neuron. (B) Quantification of seeded neurons at different time points based on reactivity to anti-HTT antibody (PHP1) (C). Examination of the lysates of seeded neurons 24 hr post-treatment by SDD-AGE detected by PHP1. HTTex1 seeds were adjusted to the amounts in the lysates loaded to determine the extent of amplification. (D & E) HTTx1 seeds and its amplified products in MESC 2.10 derived neurons are partially resistant to proteinase K (PK) digestion. Seeds (D) or neural lysates from seeded neurons (E) were treated with an increasing concentration of PK and examined by SDD-AGE and WBs using PHP2 antibody. Data are mean ± SEM; **P<0.01, ***P<0.001.

To determine whether the HTTex1 seeds indeed enter neurons, we covalently conjugated the seeds with Alexa 555 fluorophore and added to neurons. Labeled HTTex1 accumulate in the nuclei of neurons over time (Fig. 2A, left panels and B). Notably, the number of neurons with detectable labeled HTTex1 seeds reach maximal by ~12 hr of incubation but staining of the seeded neurons by the anti-HTT antibody shows ~ 4-fold increase in reactivity by 24 hr post treatment (Fig. 2A, right panels and B). One possibility is that undetectable levels of the labeled HTTx1 seeds may enter some neurons and amplify; alternatively, the newly formed assemblies released in the culture medium may be taken up by the neighboring neurons and/or transported between neurons utilizing tunneling nanotubes orchestrated by the cellular protein Rhes (Sharma and Subramaniam, 2019). Interestingly, although the HTTex1 seeds enter human neurons derived from induced pluripotent stem cells (IPSCs), they do not affect astrocytes produced from the same line even after incubation for up to 4 days in culture (Fig. 2C and D, respectively). We also examined human epithelial cell lines such as Caco-2 or mouse neuro2A cell line but found no evidence for entry and amplification (Supplementary fig. 3). Collectively, these results suggest that mutant HTTex1 seeds may preferentially enter neurons, accumulate in the nucleus and amplify into a heterogeneous mixture of aggregates.

Figure 2.

HTTex1 seeds accumulate in the nuclei of neurons. (A&B) Fluorophore-labelled (Alexa 555) HTTex1 seeds were added to human neurons for the indicated time points and processed for ICC using PHP1 antibody. Confocal images in the left panels show the labelled HTTex1 seeds and in the right panels are merged images of labelled HTTex1 and amplified species detected by PHP1 antibody. Part B is magnified neurons showing the nuclear localization of labelled HTTex1 seeds. (C) Quantification of neurons with the labelled HTTex1 in part A (left columns) and neurons labelled with anti-HTT antibody (PHP1) (right columns). (D&E) Human neurons from IPSC-derived neuronal progenitors (left panels) or IPSC-derived astrocytes (right panels) were treated with HTTex1 fibrils or sonicated fibrils for 24 hr and processed similar to part (A). Neurons were stained with PHP1 and anti-Tuj-1 and astrocytes were stained with PHP1 and anti-GFAP. Data show mean ± SEM; **P<0.01, ***P<0.001.

3.2. Mutant HTTex1 seeds produce neurotoxic assemblies

Importantly, HTTex1 seeds-triggered amplification induces robust and progressive nuclear fragmentation, which can be visualized and quantified in the DAPI- stained images (Fig. 3A and B). To confirm the toxicity of amplified HTTex1 seeds, we stained seeded neurons for active caspase-3 and find it in the majority of neurons with HTT aggregates (Fig. 3C–E). These findings indicate that HTTex1 seeds produce neurotoxic assemblies and may constitute a pathogenic component of mutant HTT.

Figure 3.

Amplification of HTTex1 seeds causes nuclear damage and caspase-3 activation. (A) For better contrast, black and white images of DAPI stained neurons treated with HTTex1 fibrils or sonicated seeds for the indicated time points are shown to display the progressive nuclear damage. (B) Quantification of DNA damage in neurons seeded with HTTex1 fibrils or seeds (C) Confocal images of neurons with active caspase-3, which was induced in neurons treated with the sonicated HTTex1 seeds. Graphs in D and E show quantification of neurons with HTT aggregates and active caspase-3, respectively. Data are expressed as mean ± SEM; ***P<0.001.

3.3. Neural entry of HTTex1 seeds deplete endogenous HTT

To better characterize the amplification of mutant HTTex1 seeds in neurons, we examined the lysates of seeded neurons by SDD-AGE followed by WBs using a panel of antibodies reactive to the PRD of HTTex1 seeds (Supplementary fig. 4A and 4B). High molecular weight assemblies (>250 kDa) are detected by monoclonal antibodies (PHP1, PHP2, PHP7, PHP8), whereas two clones (PHP9 and PHP10) react poorly (Figs. 1E, 4A). A likely possibility is that the conformations for the binding of PHP9 and PHP10 are altered in the amplified products or destroyed under denaturing conditions in SDD-AGE. Notably, examination of the seeded neural lysates with antibodies reactive to monomeric HTT species (PHP1, PHP5, PHP6, 2166) indicate depletion of endogenous full-length and N-terminal fragments of HTT (Fig. 4B, arrows, Supplementary fig. 4C). This raised the question of whether the HTTex1 seeds may recruit the endogenous HTT fragments to amplify. Sequencing the cDNA of HTT in MESC2.10 progenitors revealed 2 copies with 24Qs and 25Qs (data not shown). To confirm that HTTex1 seeds can recruit HTT fragments with normal polyQ length, we developed an in vitro amplification assay using monomeric recombinant WT HTTex1 (25Qs) as a substrate. Examined by SDD-AGE and WBs, we find that the HTTex1 seeds recruit monomeric WT HTTex1 and produce assemblies, which migrate similar to those formed in neurons (Figs. 4C and 4A, respectively). HTTex1 seeds also trigger assembly formation in neural lysates (Fig. 4D, lane 2). Moreover, the amplified products also propagate if added to fresh lysates indicating that they are biologically active (Supplementary fig. 5A). To confirm that endogenous HTT is critical, we depleted neural lysates with several anti-HTT antibodies before adding the seeds. Depletion with the anti-N17 antibodies (PHP5 and PHP6) or clones reactive to the PRD of HTTex1 (PHP1, PHP2 and PHP10) significantly reduces the amplification of HTTex1 seeds (Figure. 4D, lanes 4–6, Supplementary fig. 5B). Depletion of neural lysates with other anti-HTT antibodies such as PHP3, which binds to polyQ/polyP junction of HTT (Ko et al., 2018) or PHP8 (anti-PRD) has minor effects on seeding (Supplementary fig. 5B). These studies suggest that HTTex1 seeds may recruit the N-terminal fragments of HTT to propagate in neurons. Since the 2166 antibody, which binds to epitope outside exon1, does not react with the amplified products (Fig. 4B), it is difficult to determine whether the HTTex1 seeds also recruit full-length HTT. However, reduction in the levels of the microtubule-associated protein Tuj-1 indicates that HTTex1 seeds may promiscuously recruit other neuronal proteins, which could include full-length HTT and those interacting with it (Fig. 4B, bottom panels).

Figure 4.

HTTex1 seeds recruit endogenous HTT species to form assemblies. (A) SDD-AGE analysis followed by WBs of control neuronal lysates (C) or lysates of neurons treated with HTTex1 seeds (SD) probed with the indicated anti-HTT antibodies. Part B is SDS-PAGE of similar lysates probed with antibodies reactive to soluble HTT. Arrowheads indicate depletion of endogenous full-length and N-terminal fragments of HTT in the lysates of seeded neurons. (C) Seeding of recombinant WT HTTex1 monomers by the HTTex1 seeds (S). Products were examined by SDD-AGE and WB using PHP1 antibody. (D) HTTex1 seeds amplify in neural lysates. Lane 1 is seed alone and lane 2 is seeded neuronal lysate. The indicated antibodies below each lane were used to deplete the endogenous HTT species before seeding (lanes 3–6). The binding epitopes of PHP antibodies are shown in supplementary fig. 4A.

3.4. Antibodies to conformations in the N17 and PRD block neural entry and amplification of HTTex1 seeds

To identify the epitopes regulating the biological activity of HTTex1 in neurons, we incubated the seeds with various anti-HTTex1 antibodies (Supplementary fig. 4A) and added the seed-antibody complexes to neuronal cultures. We find that pre-incubation of HTTex1 seeds with antibodies to the N17 domain (PHP5, PHP6) and clones reactive to the PRD (PHP1, PHP2, and PHP10) blocks neural entry, amplification and nuclear damage. (Fig. 5A and B, Supplementary fig. 6). Importantly, preincubation with three other anti-PRD antibodies PHP7-PHP9, which also bind the HTTex1 seeds, does not inhibit neural seeding suggesting that distinct conformations in the PRD may be critical for seeding activity of HTTex1 (Supplementary fig. 6). Since the binding site of the protective antibodies PHP1 and PHP2 (L17 domain, Supplementary fig. 4A) have been identified (Ko et al., 2018), we mutated it to verify the importance of this epitope in neuronal seeding. Consistent with the antibody inhibition assays, sonicated fibrils of the PRD-mutated HTTex1 (HTTex1mL17) are incapable of seeding neurons and inducing nuclear damage (Fig. 5C and D). These findings suggest that conformations within the binding PHP1 and PHP2 antibodies play a prominent role in the HTTex1 seeding of neurons. Seeds pre-incubated with the N17 antibodies (PHP5, PHP6) remain bound to neurons but do not enter (Fig. 5A, right bottom panel, Supplementary fig. 6 top left panel). We predict that similar to HTTex1 fibrils (Fig. 1A) the N17-antibodies complexed to seeds may be too large to enter and/or the N-17 domain is important for steps beyond neural interaction.

Figure 5.

Antibodies to N17 and PRD of HTTex1 block neural seeding. (A) HTTex1 seeds were pre-incubated with the antibodies indicated on top of each panel, subsequently added to neurons and incubated for 24 hr. Samples were processed for ICC, stained with PHP1 and examined by a confocal microscope. Panel B is the quantification of fragmented nuclei in each condition. (C) HTTex1 seeds or those generated from HTTex1 with mutated PRD (HTTx1mL17) were added to neurons and processed as in A. Anti-N17 antibody PHP5 was used to detect the amplified products. Panel D is the quantification of fragmented nuclei in each condition. (E) MESC2.10 neurons engineered to secret control or PHP2 recombinant antibodies (Lenti-PHP2) were treated with the HTTex1 seeds, incubated for 24 hr, processed for ICC using PHP1 antibody and examined by a confocal microscope. Panel F is the quantification of fragmented nuclei in each condition. Data show mean ± SEM; ***P<0.001.

3.5. Neurons secreting anti-HTT antibodies are immune to HTTex1 seeding

Encouraged by inhibitory properties of PHP1 and PHP2 antibodies and mutation of their binding sites, we asked whether engineering neurons with recombinant antibodies provides immunity to seeding with HTTex1. Thus, MESC2.10 NPCs were transduced with a lentivirus encoding full-length PHP2 antibody. Secreted antibodies from neurons react with the HTTex1 seeds (Supplementary fig. 7). We find that neurons secreting the recombinant PHP2 are resistant to seeding by the HTTex1 since entry, amplification and nuclear damage were significantly reduced (Fig. 5E and F). These data further confirm the importance of PHP2-binding epitope of HTTex1 in neural seeding and support the potential use of recombinant antibodies as useful molecular reagents to study the impact of extracellular mutant HTT in HD pathology in animal models.

3. 6. Seeding-competent mutant HTT species accumulate in the brains of ZQ175 HD mice

To determine whether seeding-competent mutant HTT species accumulate in vivo, brain homogenates of 9-month old ZQ175 HD mice were fractionated on sucrose gradients optimized to enrich for myelin, ER/Golgi and synaptosome fractions (Mastro et al., 2020) and examined for the presence of mutant HTT by SDD-AGE and WB (Figure. 6A). Aliquots of ER/Golgi and synaptosome fractions, which had the most HTT species, were subsequently examined for entry and propagation in neurons. We find that mutant HTT assemblies in the synaptosome fractions of some animals are capable of seeding neurons (Fig. 6B); however, the seeding activity was not consistently present in the preparations from various cohorts of similar aged animals (Fig. 6C, top panels). We hypothesized that under various physiological conditions HTT may be bound to other proteins, which interfere with the neuronal entry of synaptosomal mutant HTT. Since HTTex1 seeds and those amplified in neurons are resistant to partial proteinase K (PK) digestion (Fig. 1D and E), we performed PK digestion of ER/Golgi and synaptosome fractions prior to testing in neurons. Interestingly, PK treatment significantly enhances the seeding activity of mutant HTT in the synaptosome and to some extent in the ER/Golgi fractions (Fig. 6C, bottom panels). We predict that PK treatment may expose the critical conformation of mutant HTT seeds by removing the interacting proteins; alternatively, PK treatment may directly produce seeding-competent species by cleaving mutant HTT. Similar to recombinant HTTex1, seeding by the mutant HTT assemblies from HD mice induces nuclear fragmentation (Fig. 6D). These findings suggest that seeding-competent mutant HTT species may accumulate in the brains of HD animals and are consistent with similar recent data where the mutant HTT isolated from the brains of ZQ175 HD mice seeding HEK-293 cells expressing mutant HTTex1-EGFP (Lee et al., 2020).

Figure 6.

Mutant HTT seeds accumulate in the brains of ZQ175 HD mice. (A) Fractionation of mouse brains on sucrose gradients and examination by SDD-AGE and WB. HTT was detected by PHP1 antibody. (B) Dialyzed aliquots of each fraction (50 μg) were sonicated and added into human neurons derived from MESC2.10 and incubated for 24 hr. Neurons were processed for ICC and stained with PHP1 antibody. Part C is similar fractions as in A from different cohorts of mice treated with proteinase K (PK) or untreated and processed similarly as in part B. Images were captured with a confocal microscope. (D) Quantification of seeded neurons with fragmented nuclei in each condition is shown. Data show mean ± SEM; ***P<0.001.

4. Discussion

This study was designed to investigate the biological properties of mutant HTTex1 seeds in cell models. The major findings were preferential neural entry of HTTex1 seeds, nuclear localization, amplification, and induction of neurotoxicity exemplified by depletion of HTT, nuclear damage and caspase-3 activation (Fig. 7). Our studies also implicate the N17 domain and an epitope within the PRD of HTTex1 in neural uptake and amplification. These findings and the notion that similar seeding-competent mutant HTT may exist in vivo suggest that extracellular amyloidogenic HTTex1 may contribute to the production of neurotoxic assemblies.

Figure 7.

Schematic representation of the extracellular HTTex1 journey in neurons. HTTex1 seeds potentially bind to a surface receptor to enter neurons and ultimately accumulate in the nucleus. Subsequently, the seeds may recruit the endogenous HTT fragments and other neural proteins to amplify. Amplified products induce neurotoxicity potentially by caspase-3 activation and depletion of HTT.

Recent studies suggest that HTTex1 fibrils are taken up by a number of cell lines (Masnata, et al., 2019, Lee, et al. 2020). Our studies indicate that recombinant HTTex1 seeds generated by sonication preferentially enter neurons, amplify and induce neurotoxicity, which conform with the neurodegenerative aspect of HD. Sonication, which is standard for producing seeds from mutant HTT fibrils, breaks aggregates and helps to produce smaller assemblies with increased surface area for efficient recruiting of substrates, and increases the seeding activity since the large fibrils are predominantly inert ( Kim et al., 2016, Isas et al., 2021). The generated small-size HTTex1 assemblies (Supplementary fig. 1) are likely oligomeric species capable of entering neurons and differ from large unsonicated fibrils, which bind to neurons but do not enter (Fig. 1A). As we are optimizing protocols to assemble and stabilize various oligomeric structures of HTTex1 (Isas, et al., 2021), we ultimately will be able to explore the biological properties of different assemblies of HTTex1 without sonication. This will include structural identity as well as tropism for non-neuronal cells or various neuronal lineages including medium-sized spiny neurons, which appear to degenerate first in HD patients (Bates et al., 2015). Given that HTTex1 seeds accumulate in nuclei of neurons and promote neurotoxicity (Fig. 3), characterizing their structure, size, and regulatory conformations may help to identify equivalent species in vivo and examine their role in HD progression. The accumulation of seeding-competent HTT species in the synaptosome preparations from the brains of HD mice (Fig. 6) together with the presence of HTT seeds in the CSF of HD patients (Tan et al., 2015, Lee et al., 2020) support the existence of a pathway for seeding-mediated amplification of mutant HTT assemblies in vivo. We predict that the physiologically equivalent of mutant HTT seeds in HD are early oligomeric assemblies with the propensity to move between neurons and/or are released in the extracellular space and subsequently taken up by the neighboring neurons to propagate (Sharma, M., Subramaniam S., 2019, Caron et al., 2021). However, the mechanisms of how “non-autonomous” seeding may regulate the pathogenesis of HD remains to be confirmed experimentally.

The efficient entry and amplification of HTTex1 seeds in our assays is noteworthy where up to 60% of neurons are affected within 24 hr (Fig. 1A and B). The short incubation period may be due to presence of neuron-specific receptor(s) mediating efficient entry and/or the levels of substrates including the endogenous HTT species required for amplification. HTT is abundant in neurons and reduction in the levels of full-length and N-terminal HTT fragments in the seeded neurons along with the inhibition of seeding in HTT-depleted neural lysates support a role of endogenous HTT as a substrate for the HTTex1 seeds (Fig. 4). Although amplified products have seeding capabilities (Supplementary fig. 5A), it remains to be seen whether amplification of HTTex1 seeds in neurons faithfully produces structural replicas of the original seeds thus, perpetuating the amplification of neurotoxic species. It is noteworthy that HTTex1 seeds initially accumulate in the nucleus of neurons (Fig. 2A), which may be the initial site of HTTex1 seeds amplification and toxicity. Nuclear accumulation of HTTex1 assemblies is linked to neurotoxicity and may also contribute to nuclear damage observed in our assays (Fig. 3) (DiFiglia et al., 1997, Gao et al., 2019). Since HTT is present in several cellular organelles and due to nuclear-cytoplasmic shuttling properties of HTTex1 (Saudou and Humbert, 2016), “multifocal” seeding/amplification and interruptions in the biology of other organelles may also occur. The findings that HTT levels are reduced in the seeded neurons and the amplified products are reactive to several anti-HTT antibodies support a homotypic seeding mechanism (Isas et al., 2017, Ast, et al., 2018, Pandey et al., 2018). Whether heterotypic seeding as demonstrated with α-synuclein and Tau (Jucker and Walker, 2018) also occurs with the HTTex1 seeds in neurons remains to be investigated. Amyloidogenic HTTex1 oligomers aberrantly interact with hundreds of cellular proteins, which may also be incorporated into assemblies detected in our assays (Kim et al., 2016, Wanker et al., 2019). The promiscuous interaction of HTTex1 seeds with other neuronal proteins may also contribute to the short incubation period required for assembly formation (Figs. 1 and 4).

Unraveling the robust neurotoxic aspect of extracellular HTTex1 is another novelty of our work, which might have been difficult to observe previously. The pathogenic manifestations of HTTex1 seeds in neurons included nuclear damage, caspase-3 activation and depletion of HTT, which is vital to neuronal survival. Nuclear/DNA damage is detected in the immune cells of prodromal HD patients and is recognized as a pathogenic modifier of HD (Askeland et al., 2018, Castaldo et al., 2019). The mechanism of nuclear damage in our assays remains to be investigated. However, HTTex1 seeds may sequester critical proteins including the endogenous HTT essential for maintaining the nuclear integrity (Kim et al., 2016, Gao et al., 2019, Wanker et al., 2019). HTT is a component of the ATM (ataxia telangiectasia mutated) enzymatic machinery responsible for sensing and repairing DNA damage induced by oxidative stress (Maiuri, et al., 2017). HTT also binds to a complex consisting of RNA polymerase II, CREB binding protein (CBP), ataxin 3, and the DNA repair enzyme polynucleotide-kinase −3-phosphatase (PNKP), which detects and repairs DNA damage generated during transcription elongation (Gao et al., 2019). Thus, depletion of HTT by the HTTex1 seeds may inactivate two prominent sensors of DNA damage and make the nuclei susceptible to insults. The activation of caspase-3 may further contribute to HTTex1-induced nuclear damage and apoptosis. Activated caspase-3 is detected in the brains of HD patients and cleaves HTT in human neurons (Kim et al., 2001, Khoshnan et al., 2009). Moreover, intact HTT is an inhibitor of caspase-3 and reducing its levels promotes cell death (Zhang et al., 2006). While dissecting the precise mechanism of how HTTex1 seeds induce neurotoxicity remain to be characterized, we predict that depletion of endogenous of HTT may impair various physiological pathways in the target neurons, which is manifested as nuclear crumbling and death.

Our studies also identified the critical epitopes of HTTex1 responsible for neuronal seeding. The antibody inhibition assays and mutagenesis studies indicate that conformations reactive to PHP1/PHP2 antibodies within the PRD (Ko et al., 2018) are important. The inability of some PRD-reactive monoclonal antibodies (PHP7-PHP9) to block neural seeding (Fig. 5, Supplementary. fig. 6) further reinforces the implication of novel conformations in the seeding events. Although antibodies to N17 domain also prevent seeding, the seed-antibody complexes remain avidly attached to neurons (Fig. 5, Supplementary fig. 6). We speculate that the N17 domain is not essential for binding of HTTex1 seeds to neurons but may be required for events downstream such as transport across neuronal membranes; given the high affinity of N17 for membranes (Tao et al., 2019). Consistent with our findings, recent studies indicate that N17 and PRD play a role in the uptake of mutant HTTex1 assemblies in epithelial and neuronal cells (Lee et al., 2020, Vieweg, et al., 2021). Notably, the N17 domain and PRD also regulate the neurotoxicity of mutant HTTex1 when expressed transiently in neuronal and rodent brain slice models (Shen et al., 2016). Our studies implicating the N17 and PRD in seeding-mediated neurotoxicity further supports the prominent role of these domains in the overall neurobiology of mutant HTTex1. The identification of PHP1/PHP2 binding epitope in the PRD as a critical regulator of neural entry and amplification is consistent with the findings of seeding assays performed in test tube (Ko et al., 2018) and provide a useful target to develop reagents for inhibiting the amplification of pathogenic HTTex1 assemblies. It is encouraging that engineered neurons secreting recombinant PHP2 antibodies are immune to HTTex1 seeding and subsequent neurotoxicity. This may serve as a platform to test the efficacy of other conformation-specific recombinant antibodies (Fig. 5, supplementary fig. 6) and ultimately use as novel reagents to investigate the role seeding in HD pathology in HD animal models.

In summary, our studies reveal that seeding-competent HTTex1 assemblies may preferentially enter neurons and amplify into neurotoxic assemblies. To our knowledge this is the first robust neuronal seeding assay, which facilitates the simultaneous examination of seeding activity and neurotoxicity of amyloidogenic HTT species. However, the in vitro nature of our studies limits direct application of these findings to HD pathogenesis and requires replication of these findings in animal models. It is encouraging that seeding-competent HTT species are present in HD animal models and in HD patients (Fig. 6), (Morozova et al., 2015, Ast et al., 2018, Lee et al., 2020) and these findings inspire us to extend our studies to animal models and validate any pathogenic role of extracellular HTTex1 in the manifestation of HD. The neuronal assays and reagents including the inhibitory antibodies provide a useful platform to characterize the biological properties of HTTex1 seeds from different sources, monitor disease progression in HD using fluids such as plasma or CSF, and to screen for compounds, which may prevent the amplification of neurotoxic species in vivo. The potential existence of novel receptors for the neural entry of HTTex1 species and other proteins regulating the amplification step, which are being investigated by our laboratories, may further introduce a novel pathogenic pathway and therapeutic targets for HD.

Highlights.

• Seeding-competent HTTex1 assemblies preferentially enter neurons, accumulate in the nucleus, amplify and promote neurotoxicity.

• Neuronal amplification of mutant HTTex1 seeds depletes endogenous HTT.

• The N-terminal 17 amino acids and proline-rich domain of HTTex1 regulate neural entry and amplification.

• Monoclonal antibodies to conformations in the regulatory domains of HTTex1 inhibit neural seeding.

• Seeding-competent HTT species accumulate in the synaptosome preparations from the brains of ZQ175 HD mice.

Abbreviations

HD

Huntington’s disease

HTT

huntingtin

HTTex1

huntingtin exon1

polyQ

polyglutamine

PRD

proline-rich domain

CSF

cerebrospinal fluids

IPSC

induced pluripotent stem cells

NPCs

neuronal progenitor cells

WB

Western blots

SDD-AGE

semi-denaturing detergent agarose gel electrophoresis